SYSTEM AND METHOD FOR DETERMINING THE ACTIVITY OF A PERSON LYING DOWN

• US 20120101785A1
• Filed: 04/26/2010
• Published: 04/26/2012
• Est. Priority Date: 04/24/2009
• Status: Active Grant

• Current Assignee: Commissariat a L'Energie Atomique (AES Corporation), Movea
• Original Assignee: Commissariat a L'Energie Atomique (AES Corporation), Movea SA (TDK Corporation)
• Inventors: Jallon, Pierre
• Granted Patent: US 8,898,041 B2
• Time in Patent Office: Days
• US Class Current: 703/2
• CPC Class Codes: A61B 5/1116 Determining posture transitions
• CPC Class Codes: A61B 5/1118 Determining activity level
• CPC Class Codes: A61B 5/1123 Discriminating type of movement, e.g. walking or running A61B5/1116, A61B5/112 take precedence
• CPC Class Codes: A61B 5/4094 Diagnosing or monitoring seizure diseases, e.g. epilepsy
• CPC Class Codes: A61B 5/4806 Sleep evaluation A61B5/4821 takes precedence; devices for inducing sleep A61M21/02
• CPC Class Codes: A61B 5/7264 Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems neural networks per se G06N3/00; expert systems per se G06N5/00
• CPC Class Codes: G06F 17/10 Complex mathematical operations function generation by table look-up G06F1/03; evaluation of elementary functions by calculation G06F7/544
• CPC Class Codes: G06F 2218/00 Aspects of pattern recognition specially adapted for signal processing
• CPC Class Codes: G06V 40/20 Movements or behaviour, e.g. gesture recognition recognition of facial expressions G06V40/16

Abstract

The system for determining the activity of a person lying down has at least one motion sensor (CM) having at least one measurement axis, which is provided with a fastener (MF) for firmly connecting the motion sensor (CM) to a user. The system includes a filter (FILT) for selecting, for at least one measurement axis of the motion sensor (CM), a high-frequency signal (HF) and a low-frequency signal (BF); a first calculation unit (CALC1) for calculating a first variable (x(n)) representing a temporal variation of the low-frequency signal (BF); a second calculation means (CALC2) for calculating a second variable (y(n)) comprising the Euclidean norm, along at least one measurement axis, of the high-frequency signal (HF); and an analysis unit (AN) that determines an activity of prone user as a function of time using a hidden Markov model having N states corresponding to N activities respectively. The analysis unit (AN) also combines conjoint probability density functions of the first and second variables, said probability density functions being defined for each state of the model in question; and probabilities of transitions between two successive states.

Claims

1. A system for determining the activity of a person lying down, comprising at least two processing pathways of signals at the output of at least one motion sensor substantially affixed to said person, wherein a first of said at least two processing pathways processes a first component comprising signals of low frequencies and a second processing pathway processes a second component of signals of high frequencies, said system further comprising:

first calculation means for calculating a first variable (x(n)) representing a temporal variation of said first component, for at least an axis of said motion sensor;

second calculation means for calculating a second variable (y(n)) comprising the Euclidean noun, along at least one measurement axis, of said second component; and

analysis means configured for determining an activity of said person as a function of time using a hidden Markov model having N states corresponding to N activities respectively,

said analysis means being configured for combining;

conjoint probability density functions of said first and second variables, said probability density functions being defined for each state of the model in question; and

probabilities of transitions between at least two successive states.

2. The system of claim 1, wherein the probability density functions (P(x,y)) associated with each state (i) are a function of the product of at least one first probability density function of said first variable (x) multiplied by at least one second probability density function associated with said second variable (y).

3. The system of claim 2, wherein a conjoint probability density function (P(x,y)) comprises a product of at least one probability density function (P_x,i(n)) for obtaining the first variable (x) and of at least one probability density function (P_y,i(n)) for obtaining the second variable (y), and said probability density functions (P_x,i(n) and P_y,i(n)) are defined by the following expressions:

4. The system of claim 1, wherein said analysis means are configured for determining an activity of the lying-down person as a function of time using a hidden Markov model in at most five states chosen from a rest activity, a slight agitation activity, a trembling activity, a moderate agitation activity, and a strong agitation activity.

5. The system of claim 4, wherein said analysis means are configured for determining an activity of the lying-down person from a set of predetermined pairs of values ((σ

_x[m];

σ

_y[n])) of first and second variances, defining motion classes (σ

_x, σ

_y).

6. The system of claim 5, wherein said pairs of values ((σ

_x[m];

σ

_y[n])) of first and second variances are configured for satisfying the following conditions;

7. The system of claim 6, wherein said analysis means are configured for determining the conjoint probability density function of a state of said hidden Markov model through the following equation:

8. The system of claim 7, wherein said motion classes are eighteen in number and are defined by:

σ

_x[0]=5×

10<sup>−

3</sup>, σ

_x[1]=1.8×

10<sup>−

2</sup>, σ

_x[2]=3.5×

10<sup>−

2</sup>, σ

_x[3]=5.5×

10<sup>−

2</sup>, σ

_x[4]=8×

10<sup>−

2</sup>, σ

_x[5]=1×

10<sup>−

1</sup>, σ

_y[0]=1×

10<sup>−

2</sup>, σ

_y[1]=3×

10<sup>−

2</sup> and σ

_y[2]=8×

10<sup>−

2</sup>.

9. The system of claim 8, wherein said real coefficients (α

_iState,j) are defined by the following table;

10. The system of claim 9, wherein the probabilities P(jState/iState) of said hidden Markov model, for transitions between two states representing respectively one type of posture, are such that:

11. The system of claim 1, further including display means.

12. The system of claim 1, wherein said motion sensor comprises at least one of an accelerometer, a magnetometer and a gyroscope.

13. The system of claim 1, further comprising fastening means configured for being fastened to the wrist, to the torso or to the head of the person.

14. A method for determining the activity of a person lying down, comprising at least a step of configuring two processing pathways of signals at the output of at least one motion sensor substantially affixed to said person, wherein a first of said at least two processing pathways processes a first component comprising signals of low frequencies and a second processing pathway processes a second component of signals of high frequencies, said method further comprising:

a first calculation step for calculating a first variable representing a temporal variation of said first component, for at least an axis of said motion sensor;

a second calculation step for calculating a second variable comprising the Euclidean norm, along at least one measurement axis, of said second component; and

an analysis step for determining an activity of said person as a function of time using a hidden Markov model having N states corresponding to N activities respectively,

said analysis step combining sub-steps of calculating;

conjoint probability density functions of said first and second variables, said probability density functions being defined for each state of the model in question; and

probabilities of transitions between at least two successive states.

15. The method of claim 14, wherein the conjoint probability density functions are established by determining, for each state:

first probability density functions of a first value; and

second probability density functions of a second value.